zeynepgulhan
commited on
Commit
•
27c69e9
1
Parent(s):
30bc46c
Create app.py
Browse files
app.py
ADDED
@@ -0,0 +1,60 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import time
|
2 |
+
|
3 |
+
import gradio as gr
|
4 |
+
import numpy as np
|
5 |
+
import torch
|
6 |
+
# Load model directly
|
7 |
+
from transformers import AutoModelForSequenceClassification, AutoTokenizer
|
8 |
+
|
9 |
+
device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
|
10 |
+
|
11 |
+
|
12 |
+
def get_model():
|
13 |
+
start_time = time.time()
|
14 |
+
model = AutoModelForSequenceClassification.from_pretrained("TURKCELL/gibberish-detection-model-tr")
|
15 |
+
tokenizer = AutoTokenizer.from_pretrained("TURKCELL/gibberish-detection-model-tr", do_lower_case=True,
|
16 |
+
use_fast=True)
|
17 |
+
model.to(device)
|
18 |
+
print(f'bert model loading time {time.time() - start_time}')
|
19 |
+
return tokenizer, model
|
20 |
+
|
21 |
+
|
22 |
+
tokenizer, model = get_model()
|
23 |
+
|
24 |
+
|
25 |
+
def get_result_for_one_sample(model, tokenizer, device, sample):
|
26 |
+
d = {
|
27 |
+
1: 'gibberish',
|
28 |
+
0: 'real'
|
29 |
+
}
|
30 |
+
test_sample = tokenizer([sample], padding=True, truncation=True, max_length=256, return_tensors='pt').to(device)
|
31 |
+
# test_sample
|
32 |
+
output = model(**test_sample)
|
33 |
+
y_pred = np.argmax(output.logits.detach().to('cpu').numpy(), axis=1)
|
34 |
+
return d[y_pred[0]]
|
35 |
+
|
36 |
+
|
37 |
+
def process_sentence_with_bert(sentence):
|
38 |
+
print('processing text with bert')
|
39 |
+
start = time.time()
|
40 |
+
result = get_result_for_one_sample(model, tokenizer, device,
|
41 |
+
sentence) # Bu fonksiyonun implementasyonunu sağlamalısınız.
|
42 |
+
print(f'bert processing time {time.time() - start}')
|
43 |
+
return result
|
44 |
+
|
45 |
+
|
46 |
+
def classify_gibberish(sentence, ignore_words_file):
|
47 |
+
# ignore_words_file işlenmesi gerekiyor. Gradio dosya yükleme ile ilgili bir örneği aşağıda bulabilirsiniz.
|
48 |
+
result = process_sentence_with_bert(sentence)
|
49 |
+
return result
|
50 |
+
|
51 |
+
|
52 |
+
iface = gr.Interface(fn=classify_gibberish,
|
53 |
+
inputs=[gr.Textbox(lines=2, placeholder="Enter Sentence Here..."),
|
54 |
+
gr.File(label="Upload Ignore Words File")],
|
55 |
+
outputs=gr.Textbox(label="Gibberish Detection Result"),
|
56 |
+
title="Simple Gibberish Text Detection For Turkish",
|
57 |
+
description="""Simple gibberish text detection given text like
|
58 |
+
adsfdnsfnıunf
|
59 |
+
sasdlsöefls.""")
|
60 |
+
iface.launch()
|